AI's Energy Appetite Fuels Data Centre Sustainability Debate

Data centres stand at the heart of the AI revolution, but the rising energy demands linked to AI now challenge the very sustainability standards these digital hubs must uphold.

Arm, a semiconductor and software design company, has brought sharp focus to the growing tension between AI advancement and environmental responsibility.

Their findings highlight how energy-hungry data centres are testing infrastructure capacity and climate commitments, while skills shortages threaten to stall progress.

Powering AI: The data centre challenge

AI systems require vast amounts of computational power, especially during model training and inference, processes that must be supported by energy-intensive infrastructure.

According to Arm, global data centres already consume 460 terawatt-hours (TWh) of electricity each year.

Dr Vanessa Just, Founder and Chief Executive Officer of JUS.TECH GmbH, a sustainability consultancy, draws a stark comparison: “Today’s data centres already consume lots of power: Globally 460 terawatt-hours (TWh) of electricity are needed annually.

"That’s equivalent to the entire country of Germany.”

This level of consumption not only raises environmental concerns but also introduces challenges for electrical grids.

In the US, data centre electricity demand is projected to triple, from 2.5% of the national total in 2022 to 7% by 2030, equating to approximately 390 TWh.

That figure mirrors the consumption of 40 million American households.

In regions like Virginia and Ireland, the strain is already visible.

Grid congestion has led to delays in building new facilities, with some utility companies implementing restrictions on new connections.

Dr Nicole Höher, Project Manager for Sustainability & Digitalisation at JUS.TECH GmbH, warns: “Without significant infrastructure investment, the risk of grid instability and supply constraints grows.”

To navigate these issues, Arm advocates for cross-sector collaboration between infrastructure developers, technology firms and energy providers.

The aim is to ensure AI’s energy appetite does not compromise global climate strategies.

Engineering sustainability through hardware innovation

While software improvements play a part, Arm places strong emphasis on hardware’s role in curbing energy use.

Efficient chip design offers a key opportunity to make data centre operations more sustainable.

AWS Graviton processors, built on Arm technology, are cited as one such solution.

They can lower the carbon intensity of cloud-based AI workloads by up to 67% compared to traditional x86 processors.

In mobile and edge computing, Arm-based accelerators achieve 50 to 80% energy savings over general-purpose graphics processing units (GPUs).

For data centres specifically, Arm Neoverse processors show promise, potentially reducing server rack energy consumption by up to 40%.

Yet despite these innovations, the urgency remains.

“Without proactive measures, AI-driven energy consumption could push the world further off track from climate targets, with projections indicating an increase of over 2°C in global temperatures; breaching the recommendations to limit the rise to a much safer 1.5°C,” says Maureen McDonagh, Head of Sustainability at Arm.

In parallel, concerns about electronic waste are rising.

Arm cites estimates from The Register that suggest Gen AI alone could produce an additional 2.5 million tonnes of e-waste annually by 2030 unless mitigation strategies are adopted.

Building skills to match AI investment

Energy use is not the only sustainability-related obstacle.

The industry also faces a knowledge gap. While 75% of businesses have adopted AI tools, only a third of employees have had any AI training in the past year.

This disconnect limits the return on technological investment and impedes effective integration.

According to the Arm AI Readiness Index, 34% of organisations say they are short on AI talent, while 39% report no programmes to upskill their existing workforce.

Communication is lacking, too. Just 15% of US employees believe their employer has outlined a clear AI strategy and only 11% feel “very prepared” to use AI in their roles.

Mark Hinkle, Chief Executive Officer and Founder of Peripety Labs, says: “Corporate leaders risk underutilising their AI investments when the workforce isn’t on board.

"When employees aren’t empowered to use AI, those expensive platforms and algorithms can turn into costly shelfware, delivering only a fraction of the promised productivity gains. AI adoption isn’t just a digital transformation – it’s a people transformation.”

Organisations are experimenting with various training strategies.

One company explains: “Our organisation offers a comprehensive AI training programme that includes workshops, online courses and practical projects.”

Another adds: “Selected personnel from each department undergo intensive Python development education twice a week for one month, with certification exams conducted upon completion.”

Arm backs this approach and has gone further by investing in the future workforce.

Through the CASCADE Centre at the University of Cambridge, Arm supports 15 PhD students researching processor designs for AI.

Khaled Benkrid, Senior Director of Education and Research at Arm, explains: “In a world where humans and machines are working together more than ever, the ability to build and use AI tools effectively is becoming a fundamental skill.

“Companies must invest in training programmes that help employees understand AI's capabilities and applications,” says Khaled.

“Furthermore, the future workforce will need to combine human ingenuity with new and emerging AI technologies; going beyond just the technical skills.”

Explore the latest edition of Data Centre Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today.

Subscribe to our Weekly Newsletter for the latest insights, trends and expert analysis.

Data Centre Magazine is a BizClik brand